Back to skill

Security audit

utilitarianism

Security checks for vulnerabilities and agentic risk

Overview

This skill provides a structured way to compare benefits and harms without requesting credentials or control of external systems.

Use this as a reasoning aid. Verify numerical examples and assumptions before relying on its conclusions for important decisions.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 88)May include surrounding context.

md
|---|---|
| [D] Omitting identifiable affected parties. | Omitting a cost-bearing group biases the conclusion — not neutral. |
| [D] Claiming net positive welfare without stating the conversion rate. | Every utilitarian analysis embeds a conversion rate; make it explicit. |
| [D] Treating welfare-maximizing as the final answer without checking rights constraints. | Aggregate optimization does not resolve whether non-overridable rights are violated. |
| [D] Using uncertain welfare estimates as if precise. | State uncertainty range; conduct sensitivity analysis. |
| [D] Ignoring second-order welfare effects. | Trace consequence chain at least two orders before concluding. |
| [D] Conflating "most people benefit" with total welfare maximization. | Total welfare = sum of magnitudes, not count of beneficiaries. |